We have a new fastest system for both UHD and FHD- AMD5950x with RX-6900XT. Thanks to Léonce for submitting. Link to the results on the first page.

We have a new fastest system for both UHD and FHD- AMD5950x with RX-6900XT. Thanks to Léonce for submitting. Link to the results on the first page.

<3

that is one beast of a computer!

Yes, it is really something, it's true. Quite impressive and so pleasant to work with.

And a beast build for Magix Vegas !

Thanks to this forum, our little community and all of our messages trying to understand, days after days, what is the best setup to feed our lovely vegas.

Just for your information and add my two cents in the "Smart access Memory", my old computer used to render an HD file at 50 frames / second.

New one out of the box : 155 frames / second.

New one with SAM enable and D.C.O.P for the ram : 200 frames / second.

I forgot about this topic until it was mentioned in another thread. I tried this benchmark again with current windows updates and drivers, it worked fine, but do you guys know if it's a known fact that Vegas 18 and below is not compatible with 3000 series NVENC encoding? It works but is slow and speeds are highly variable?

VP19 4k Nvenc - 1m15s 1080p Nvenc - 45s

VP18 4k Nvenc - 3m30s,3m50s,4m15s 1080P Nvenc 2m13,2m17s,1m31, 1m42s

VP18 4k Software - 1m40s 1080p software - 1m06s

VP17 4k Nvenc - 4m14s,3m41s 1080p Nvenc - 1m01s,52s

I guess this is a selling feature of VP19 even if timeline playback was slower when I tested with it on release

Hi Todd, while I've been recommending people try this project for diagnostic purposes it really is most useful as a benchmark. With enough data we can interrogate the results in different ways, including to answer the question you posed as to support for 3XXX NVIDIA cards and NVENC renders.

The problem is not enough data at present to go beyond anecdotes. You can help with that by submitting the reporting form for your renders for VP 19, 18 and 17 to the database (use the Google form) and try to encourage other NVIDIA users to do the same.

do you guys know if it's a known fact that Vegas 18 and below is not compatible with 3000 series NVENC encoding?

A while back I got a Nvidia 3060 Ti for $400 from Best Buy online, tested it using V18 with the "Sample Project 4K" benchmarks, and got UHD scores (1:20) slightly better than my AMD VEGA 64 LQ (1:34). However, I decided to sell it while resell prices are high & wait for something better... The only problem I had with V18 & NVENC encoding was that I could only open & render using (3) instances of Vegas whereas with AMD VCE I think the only limitation is system RAM...

BTW, I'm also thinking of selling my AMD VEGA 64 LQ while its market value is $800+ & my VEGA 56 @ $700+... That would give me $1500 towards better GPUs like the AMD 6900XT when prices & availability stabilize. In the meantime, I can get-by with slower QSV renders since I use (1) PC for editing & (1) PC for rendering...

@RogerS There's something unusual going on, @TheRhino 's benchmarks don't show any peculiarities with VP18, but I have heard others talk about slow NVENC with their 3000 series Nvidia cards. Requires further investigation, Maybe a plugin is responsible, I would not like to submit data that is not typical of my CPU and GPU if my VP18 is not working correctly

Please submit your data and encourage others who "talk about slow NVENC" to do the same. The talk I saw was early in the 3XXX release cycle, before studio drivers were available. I thought it was resolved.

If you later find your data is faulty due to some confounding issue, just message me or post here and I'll delete it from the benchmark table (or revise it with new data.)

@RogerS I can do VP19, not VP18, as the figures would be misleading in helping someone buy a machine for Vegas

Timestamp
User Name        Todd-b
Computer brand/model    home build
Version of Vegas Pro    VP19
Build of Vegas Pro    381
CPU Manufacturer    AMD
CPU Model        Ryzen5900x
CPU Cores        12
Main GPU Mfg.        Nvidia
GPU Model        rtx3070
Integrated GPU?        no            
Overclocked?        no

    
Render frame size    FHD    
Hardware Graphics Acceleration    yes
Render File Type    Magix AVC
Encode Mode        Nvenc
Render time (minutes and seconds)    0:44s
Average Frames Per Second for Playback    10
Comments

    
Render frame size    FHD    
Hardware Graphics Acceleration    yes
Render File Type    Magix AVC
Encode Mode        Software
Render time (minutes and seconds)    48s
Average Frames Per Second for Playback    10
Comments

    
Render frame size    UHD
Hardware Graphics Acceleration    yes
Render File Type    Magix AVC
Encode Mode        Nvenc
Render time (minutes and seconds)    1:13
Average Frames Per Second for Playback    10
Comments

    
Render frame size    UHD
Hardware Graphics Acceleration    yes
Render File Type    Magix AVC
Encode Mode        Software
Render time (minutes and seconds)    1:36
Average Frames Per Second for Playback    10
Comments

Thanks for filling out the form- you'll see the results on the results page. First impression is looks strong and in line with what others are getting for both CPU and GPU renders. Hopefully we can get more submissions with newer NVIDIA and AMD GPUs.

The UHD software encode seems a bit fast, If I read the benchmark chart correctly I'm in equal 2nd place for fastest CPU render? That seems a bit unusual unless VP19 is faster and most of the comparisons are with older versions. My computer is very average with CPU limited to 170watts. I will check that again tomorrow

It's a small data set but you are 5th fastest CPU render and 7 seconds off a Ryzen 9-5950x which seems plausible.

I have not seen performance differences between VP 18 527 vs VP 19 as of yet (but very limited data; I did tests and Howard did too, but he tracks his results separately.)

@RogerS Cudos on what you've done with getting submissions via Google Forms and putting them online into Google Sheets. My new system build is looking nearly finalized with a 6900xt... will submit something run manually later today if that's OK.

That would be wonderful, look forward to 6900xt data in particular to see what difference it makes.

Edit: The i9-11900K with RX-6900XT and VCE is now the combo to beat!

@RogerS @JN- Hi , I've been trying this benchmark project, playback on the selected area is about 10-17fps

the fastest i can get is from the default render settings for 1080 HD

0.43secs

2160 UHD is 1.08secs

I don't know how to create the data list that people have been posting?

Hi Gid, JN- isn't on this forum anymore.

If you go back to page 1 of this thread you can see a new online form you can use to upload your results.

10-17FPS is extremely fast for that section. Try to play it back a few times and just average the performance. Turning dynamic ram preview to 0MB gives a more realistic figure (it is a buffer). Then set it back to whatever number you normally use for dynamic ram preview before you do the render.

@RogerS Thanks,

Setting dynamic ram preview to 0MB improved playback a fraction, but when the selection is set to region 1 it stutters for a moment as it loops & gets started playing again, it doesn't play properly until 1/3 of the way through, 10-17fps like i say when i do just that region,

If i play the whole project as it passes over region 1 & is in full playing mode i get 12-20fps,

Yesterday when i did the ones in the pics in the comment above, the PC had been on for 14hrs+, I was told when i bought it it runs hot & it does like to work better when it's warmed up, i ran this test a dozen times, I tried changing a few settings on my GPU control panel, no improvement which each render,

I decided to wait & try it again in the morning when the PC had cooled over night, the render times were fractionally slower, 48secs compared to the 43secs i was getting the night before,

I ran it again an hr or so later after the PC had warmed up a bit, 46secs 🤷‍♂️😕 then i remembered i'd turned dynamic ram preview to 0MB, so i gave it the 15% i normally have it set at for building dynamic ram preview, render times went back down to 43secs & 1.07secs just like the night before.

I have 256GB RAM so the setting looks like this

I'll do the online form later 👍

It's supposed to worsen playback on this very challenging section. Include the stuttering part in your average framerate.

Thanks for your report and if you fill out the form I don't have to enter it all manually : )

It will also be a good reference as drivers and Vegas are updated to see if performance improves or regresses from time to time.

Thanks, ok & yeah i understand the reason for that selection,

One thing that's always puzzled me is however many times i play something, it never seems to remember, it plays a little better after the first couple of times but seems to reach a point where there's no improvement,

I just went to the shop, i was gone ½hr-ish, i left this benchmark test looping, when i returned it was still only playing 12-15-25fps at different parts of the clip, you'd think after the first dozen times it'd remember what it's playing,

Is there no cache type thing in Vegas?

Dynamic ram preview is the cache and when it fills up it overwrites itself so no more learning can take place. At least that's how I understand it- it's a bit mysterious and due for an upgrade at some point.

Thanks, that's why i have some Dynamic ram saved & not set to 0 to do that ram preview 👍, yeah an upgrade would be good, a lot of my other software plays through once then it's good to go on the second time around, like you say it does need to update if i make any changes but it's automatic.